Forecasting photovoltaic power using bagging feed-forward neural network

This paper presents a forecast model of the active power of a photovoltaic (PV) power generation system. In this model, a feed-forward neural network (FNN) is combined with bootstrap aggregation techniques using the Box–Cox transformation, seasonal and trend decomposition using Loess, and a moving b...

Descripción completa

Detalles Bibliográficos
Autores: Zarate, Juan, Palumbo Fernández, Mariana|||0000-0002-9157-0943, Torres Seroa da Motta, Ana Lúcia, Grados Gamarra, Juan Herber
Tipo de recurso: artículo
Fecha de publicación:2020
País:España
Institución:Universitat Politècnica de Catalunya (UPC)
Repositorio:UPCommons. Portal del coneixement obert de la UPC
Idioma:inglés
OAI Identifier:oai:upcommons.upc.edu:2117/345501
Acceso en línea:https://hdl.handle.net/2117/345501
Access Level:acceso abierto
Palabra clave:Photovoltaic power generation
Bootstrap
Demand forecasting
Energy consumption
Feed-forward neural networks
Photovoltaic systems
Energia solar fotovoltaica
Àrees temàtiques de la UPC::Energies::Energia solar tèrmica
Descripción
Sumario:This paper presents a forecast model of the active power of a photovoltaic (PV) power generation system. In this model, a feed-forward neural network (FNN) is combined with bootstrap aggregation techniques using the Box–Cox transformation, seasonal and trend decomposition using Loess, and a moving block bootstrap (MBB) technique. An analysis is conducted using the data provided by the active power of the PV power generation system; the data are collected every 30 min for 12 months. The FNN method combined with MBB techniques consistently outperformed the original FNN in terms of forecasting accuracy based on the root mean squared error, on the forecast from one day of anticipation. The results are statistically significant as demonstrated through the Ljung–Box test, which verifies that the forecast errors are not correlated, thereby validating the proposed model.